Using the caBIO Portal Freestyle Lexical Mine Tool to Find Genes Associated with a Disease

Because the Freestyle Lexical Mine suggests caBIO terms that match the characters you have entered, it is relatively easy to find a disease search term that matches your desired disease concept. If you would like to search for any disease term that contains a string (set of alphanumeric characters) use the special character "*" which will retrieve results will have zero or more characters in place of the asterisk. For example, "ovarian serous adenocarcinoma" would only return objects that are associated with this exact gene term, but "*ovarian*" will retrieve objects that are associated with any term that contains the string "ovarian."

After entering a search term (1), click the Submit button (2) to retrieve results. Although you may limit your search by clicking on the more options ... link, this is not required (3).

The caBIO Portal Freestyle Lexical Mine will group the retrieved objects that match your disease term of interest by type. To view genes that are associated with your disease term of interest, click the Evidence tab at the top of the page (Figure 5.2, 1). Each row in the Evidence results table is a truncated view of an Evidence object (i.e., not all attributes and methods are shown here), whereas the columns show the object's attributes: class and identifier (Figure 5.2, Class/Id), the evidence of the gene-disease association (Sentence), whether the evidence was collected from experiments involving cell lines (_Figure 5.2, Cellline Status), whether the evidence is negative (i.e., gene X is not associated with disease Y; Figure 5.2, Negation Status), the PubMed identifier for the abstract from which the evidence was extracted (Figure 5.2, Pubmed Id), and whether the sentence supported the candidate gene-disease association (finished), refuted the association (no_fact), etc. (Figure 5.2, Sentence Status). For additional information on the evidence, annotations, and the process by which they were generated please see Cancer Gene Index #Data, Metadata, and Annotations and #Creation of the Cancer Gene Index.

Warning! If you do not want to spend time navigating through the caBIO object model for candidate gene-disease associations that were found to be false positives, unclear, or redundant to other data, select only Evidence objects where the Sentence Status is finished.

Figure 5.2. Disease Term Search Results for "Ovarian Serous Carcinoma" using the caBIO Portal Freestyle Lexical Mine

To discover which gene is associated with each piece of evidence, click on the Class/Id link for the desired object. This will open the full Evidence type object. Scroll over to the right, and click on getGeneFunctionAssociationCollection method link (Figure 5.3, 1) to view the Gene Disease Association type object. This object has a role attribute that contains one or more #Role Codes or #Role Details that describe the nature of the gene-disease relationship, as well as a notation that the Cancer Gene Index is the source of these data.

Note: A single piece of evidence may have multiple Role Codes and Role Details describing the gene-disease association, and the evidence may also describe gene-compound associations. Thus, after clicking the getGeneFunctionAssociationCollection link, you may see multiple retrieved objects of type gov.nih.nci.cabio.domain.GeneDiseaseAssociation and even multiple object retrieved records of type gov.nih.nci.cabio.domain.GeneAgentAssociation.

Click on the getGene link (Figure 5.3, 2) to access the related Gene object (Figure 5.3, bottom panel). This Gene object contains the full name and HUGO Gene Symbol in the fullName and hugoSymbol columns (Figure 5.3, 3), for example, for the gene associated with the disease of interest and a specific piece of evidence.

To uncover additional genes associated with the disease term of interest, navigate back to the evidence page (Figure 5.2) and repeat this process.

Figure 5.3. Navigating the caBIO Object Model to Find Genes Associated with the "Ovarian Serous Carcinoma" Disease Term Search Using the caBIO Portal Freestyle Lexical Mine.

It is possible to view disease ontologies by navigating through the object model to records for objects in the class Disease Ontology, gov.nih.nci.cabio.domain.DiseaseOntology. You may get to these records from either the Disease Ontology tab from the main search result page or by clicking through the model starting from retrieved Evidence type objects. To view related disease concepts from the Disease Ontology tab, click on the tab and select the any hyperlink in the Class/Id column. This will reveal a Disease Ontology type object. If you would like to view parent or child disease concepts linked with an evidence object, select the getGeneFunctionAssociationCollection link on the Evidence object page (Figure 5.3, 1) and then click the getDiseaseOntology link in the GeneDiseaseAssociation object record (Figure 5.3, Middle panel, at right). Once you have pulled up a Disease Ontology record, you can find parent disease concepts by scrolling to the right and selecting the getParentDiseaseOntologyRelationshipCollection link; child disease concepts can be accessed by clicking on the getParentDiseaseOntologyRelationshipCollection link.

Warning! caBIO may not always contain the most up-to-date disease ontology data from the NCI Thesaurus. If the disease concept of interest has neither parent nor child concepts, search the NCI Thesaurus using your disease term or EVS ID listed in the EVS ID column of the Disease Ontology object.

Text